Chemomechanical mapping of ligand-receptor binding kinetics on cells

Abstract

The binding kinetics between cell surface receptors and extracellular biomolecules is critical to all intracellular and intercellular activity. Modeling and prediction of receptor-mediated cell functions are facilitated by measurement of the binding properties on whole cells, ideally indicating the subcellular locations or cytoskeletal associations that may affect the function of bound receptors. This dual need is particularly acute vis à vis ligand engineering and clinical applications of antibodies to neutralize pathological processes. Here, we map individual receptors and determine whole-cell binding kinetics by means of functionalized force imaging, enabled by scanning probe microscopy and molecular force spectroscopy of intact cells with biomolecule-conjugated mechanical probes. We quantify the number, distribution, and association/dissociation rate constants of vascular endothelial growth factor receptor-2 with respect to a monoclonal antibody on both living and fixed human microvascular endothelial cells. This general approach to direct receptor imaging simultaneously quantifies both the binding kinetics and the nonuniform distribution of these receptors with respect to the underlying cytoskeleton, providing spatiotemporal visualization of cell surface dynamics that regulate receptor-mediated behavior.

Fig. 1.

Chemomechanical imaging of individual VEGFR2 receptors on fixed HUVEC surfaces. (A) Phase image of cell body and periphery. (B) Recognition image over indicated area in A shows strong binding events between the anti-VEGFR2 functionalized probe and the cell surface as discrete, dark spots (e.g., circled) that are ostensibly VEGFR2. (C and D) The capacity to block these binding events by addition of 5 μg/ml soluble anti-VEGFR2 supports this binding specificity between the probe and VEGFR2, as the number of recognition events decreases with time postblocking of 12 (C) and 60 (D) min. (F–I) Corresponding height images (F–H) indicate the position of four cytoskeletal bundles (shown only in H for clarity, as blue bands reconstructed from height traces such as I). These stiff, subsurface bundles can be correlated with receptor position and show that cell topography is stable over this time scale. Bundles are identified as F-actin through structural correlations between fluorescence optical images (FITC–phalloidin stains F-actin) and conventional AFM height images (E). (Scale bars: 10 μm, A and E; 500 nm, B–D and F–I.) Imaging in Hepes buffer at 27°C at scan rates of 10 μm/sec (A) or 1 μm/sec (B–D and F–I).

Fig. 2.

Confirmation of anti-VEGFR2 binding specificity to VEGFR2 receptors on cell surfaces. (A and B) After imaging fixed HUVECs with anti-VEGFR2-functionalized probe [phase image (A) and recognition image in region of interest (B)], a soluble antibody against a different HUVEC receptor was added. (C) Thirty minutes after addition of anti-CD31 (5 μg/ml), no competitive blocking of recognition events was observed, indicating that recognition events are specific between the cell VEGFR2 and probe-bound anti-VEGFR2. (Inset) Corresponding height image indicates the position of cytoskeletal bundles beneath the cell membrane; as in Fig. 1, bundle edges (white lines) were constructed from height traces such as E. (D) Output voltage scale for B and C shows recognition signal compared with background in a line scan over a region including three binding events (receptors). B–E show that receptors are nonuniformly distributed near cytoskeletal bundles beneath the plasma membrane. (Scale bars: 10 μm, white; 500 nm, black.) Scan rates: 10 μm/sec, A; 1 μm/sec, B and C.

Fig. 3.

Force spectroscopy analysis on fixed HUVECs to extract k_off. (A) Representative force–displacement profile for specific ligand–receptor unbinding between the anti-VEGFR2 probe and imaged receptor recognition sites. (B) Distribution of >600 rupture forces measured at recognition sites, indicating maxima of 33 and 64 pN. (C and D) Representative profiles at >400 nonrecognition sites on the cell surface indicate a nonspecific rupture force level of 13 pN. Effective loading rate: 11.7 nN/sec. Bond lifetime τ in A is proportional to the binding displacement and is used to calculate binding kinetics (see Materials and Methods).

Fig. 4.

Receptor imaging during competitive inhibition. Receptor recognition sites in images such as Fig. 1B decrease with time after addition of soluble anti-VEGFR2 (5 μg/ml, 27°C). As the number of observable sites decreases during blocking, the number of receptors bound by soluble antibodies correspondingly increases (red filled circles). Kinetic constants k_on and K_D can be determined by application of a binding kinetic model for which k_off is assumed from independent force spectroscopy experiments (blue line), or by a least-squares best fit to the experimental data (pink line). See SI Text for detailed calculation of binding kinetics.

Fig. 5.

Receptor imaging on living HUVEC surface. (A) Portion of living cell imaged with anti-VEGFR2-functionalized probe in magnetic AC mode at 27°C, phase image. (B and C) Ligand–receptor binding results in punctate image contrast (circled regions are subset of observed receptors) in phase lag images that is competitively inhibited with soluble anti-VEGFR2 antibody (data not shown). Time lapse between B and C is 30 min. Note mechanical contrast and displacement of underlying cytoskeletal actin (normal to arrow) over this time scale. These images indicate 1.32 ± 0.44 × 10⁵ receptors per cell (n = 6). Scan rates: 10 μm/sec, A; 1 μm/sec, B and C. (Scale bars: 10 μm, A; 500 nm, B and C.)